Assessment driven process modeling for software process improvement

Abstract

Software process improvement (SPI) is used to develop processes to meet more effectively the software organizationpsilas business goals. Improvement opportunities can be exposed by conducting an assessment. A disciplined process assessment evaluates organizationpsilas processes against a process assessment model, which usually includes good software practices as indicators. Many benefits of SPI initiatives have been reported but some improvement efforts have failed, too. Our aim is to increase the probability to success by integrating software process modeling with assessments. A combined approach is known to provide more accurate process ratings and higher quality process models. In this study we have revised the approach by extending the scope of modeling further. Assessment Driven Process Modeling for SPI uses assessment evidence to create a descriptive process model of the assessed processes. The descriptive model is revised into a prescriptive process model, which illustrates an organizationpsilas processes after the improvements. The prescriptive model is created using a process library that is based on the indicators of the assessment model. Modeling during assessment is driven by both process performance and process capability indicators.